机器学习——KNN

导入类库

 import numpy as np
 from sklearn.neighbors import KNeighborsClassifier
 from sklearn.model_selection import train_test_split
 from sklearn.preprocessing import StandardScaler
 from sklearn.linear_model import LinearRegression
 from sklearn.metrics import r2_score
 from sklearn.datasets import load_iris
 import matplotlib.pyplot as plt
 import pandas as pd
 import seaborn as sns

# 熵增益
# 熵越大，信息量越大，蕴含的不确定性越大

KNN
1.计算待预测值到所有点的距离
2.对所有距离排序
3.找出前K个样本里面类别最多的类，作为待预测值的类别

代码


 A = np.array([[1, 1], [1, 1.5], [0.5, 1.5]])
 B = np.array([[3.0, 3.0], [3.0, 3.5], [2.8, 3.1]])
 
 def knn_pre_norm(point):
     a_len = np.linalg.norm(point - A, axis=1)
     b_len = np.linalg.norm(point - B, axis=1)
     print(a_len.min())
     print(b_len.min())
 
 def knn_predict_rev(point):
     X = np.array([[1, 1], [1, 1.5], [0.5, 1.5], [3.0, 3.0], [3.0, 3.5], [2.8, 3.1]])
     Y = np.array([0, 0, 0, 1, 1, 1])
 
     knn = KNeighborsClassifier(n_neighbors=2)
     knn.fit(X, Y)
 
     print(knn.predict(np.array([[1.0, 3.0]])))
 
 def iris_linear():
     # 加载iris数据
     li = load_iris()
     # 散点图
     # plt.scatter(li.data[:, 0], li.data[:, 1], c=li.target)
     # plt.scatter(li.data[:, 2], li.data[:, 3], c=li.target)
     # plt.show()
     # 分割测试集和训练集，测试集占整个数据集的比例是0.25
     x_train, x_test, y_train, y_test = train_test_split(li.data, li.target, test_size=0.25)
     # 创建KNN分类，使用最少5个邻居作为类别判断标准
     knn = KNeighborsClassifier(n_neighbors=5)
     # 训练数据
     knn.fit(x_train, y_train)
     # 预测测试集
     # print(knn.predict(x_test))
     # 预测np.array([[6.3, 3, 5.2, 2.3]])
     print(knn.predict(np.array([[6.3, 3, 5.2, 2.3]])))
     # 预测np.array([[6.3, 3, 5.2, 2.3]])所属各个类别的概率
     print(knn.predict_proba(np.array([[6.3, 3, 5.2, 2.3]])))
 
 if __name__ == '__main__':
     # knn_predict_rev(None)
     # knn_pre_norm(np.array([2.3,2.3]))
     iris_linear()